Running a design sprint at Lyst
The winding road to product-market fit
In Q4 2019 the Lyst CX tribe decided to shake up the normal processes in favour of a bolder approach. We decided to undertake a design sprint in an effort to explore where we could take our product in the future.
The goal of running a design sprint at Lyst was to find a highly retentive product with strong product-market fit. The idea was to identify possible opportunities and quickly test solutions with real users using prototypes.
It was also an opportunity to solve user problems without the overheads or constraints of pre-existing solutions. It enabled us to think more freely and examine how we would solve problems in an ideal environment.
Due to the scope of what we were trying to achieve we changed the standard GV approach to allow more time to design, test, refine (and throw out) prototypes. This was achieved by extending the length of the sprint to three weeks. Week one was similar to GV’s, focusing on ideation and discussion. The second and third weeks were focused on prototype refinement and iteration.
The high level approach was:
- Understand: Understand our users and the problem space
- Ideate: Come up with ideas for how we might solve our users problems
- Sketch: Low fidelity sketches and exploration of select ideas high impact and lower cost
- Design: High fidelity prototypes to be created
- Test and learn: Test prototypes with real users (using usertesting.com and recruitment)
- Refine: Make changes to prototypes based on feedback
Repeat steps 4–6 until we judge our idea to be validated.
7. Build MVP
Understanding and prep
Preparation consisted of gaining an understanding of the specific problems we were trying to solve and who we were solving them for.
Lyst is obviously focused on helping users to find items of clothing they are looking for. Within this purview, there were two areas we wanted to explore:
- Helping users find clothing when they knew exactly what they wanted (e.g. I want a pair of Balenciaga Triple S trainers in white), and
- Helping users find clothing when they only vaguely knew what they were looking for (e.g. I want a jumper in my style).
Two teams were formed and assigned one of the two above problem areas. I was on the team focused on problem 2 so that is what we will be looking at here.
A number of personas were created to help steer the discussion. These personas were indicative of behavioural patterns that had been observed through surveys and interviews in the preceding six months. They were not specifically Lyst users. It was also clear that some of the personas were more suited to our particular problem space than others.
Kickoff workshop and ideation
With preparation complete, we kicked off with a day long workshop. Representatives from multiple areas of the business and various disciplines attended.
The morning section of the workshop focused on generating ideas by examining our problem space through the lens of our personas. We looked at problems and opportunities in each persona and constructed “how might we” statements to hypothesise how to solve them. This exercise was completed for each persona.
In the afternoon, we split the entire group into pairs. These pairs chose two ideas that were generated in the morning and fleshed them out using provided customer journey templates. These were then presented to the wider group.
As a group we plotted these more refined ideas in an opportunity/cost matrix. Ideas were then chosen that had a big opportunity with the least cost. In our case, it was decided to pursue the idea of a “personalised store.”
Finally, we had to better understand and explore our proposed idea. Designers sketched out wireframes and flows and met with their teams to discuss, refine and give direction.
Part 1: The personalised store
The basic premise behind the personalised store was that we would take the common features in popular ecommerce fashion apps and add a layer of personalisation atop these.
Initially the focus was placed on new items, but we moved to being more “store like” as we thought users would more easily understand the product.
This more store-like approach meant copying the common features and patterns in fashion ecomm app (e.g. new in, sales, editorial, category navigation, wishlist) space and personalising these.
Iterations changed rapidly. Much of the feedback from users was centred on communicating to users that their experience was being personalised.
The majority of personalisation in this prototype took place in the “home” and “browse” screens:
- In the “home screen”, we had a number of different home screens for users who fell into different styles (e.g. street style, minimal, luxury) and we would show a different home screen to each
- In the “browse” screen there was a switch which would take users to feeds of products from brands they follow after they tap into a category.
As we began scoping for an MVP of the personalised store, we felt the product was too big (in terms of dev and features) and not specific enough in the problems it was trying to solve. In addition we were trying to solve problems for a number of personas, which we felt meant not solving any of them to a satisfactory degree.
Part 2: What’s new 10x
We took a step back and started again. This time we were laser focused on Olivia and her needs. We conducted further interviews with Olivia-types (recruited through surveys).
Through these interviews we observed three key findings:
- Olivias browsed daily, with a focus on new arrivals
- Olivias browsed from a number of different retailers/brands
- Olivias bought 70% of their clothing through the above journeys
We felt there was an opportunity to streamline Olivia’s experience by combining the “new in” area of all of her favourite brands and retailers.
The prototypes based on this idea tested well with users. Many interviewees said they would add this product to their daily rotation of fashion apps and could see themselves using solely this application.
It was decided that the “What’s new” prototype is validated to the point that it is worth further testing with more users. The prototype will be broken down into an MVP, shipped and iterated upon. Undoubtedly it will be unrecognisable if it ever makes it out into the wild. But the design sprint served its purpose in helping steer us in a radical new direction.